Arden University, Birmingham, UK; Department of Psychology, University of Warwick, Coventry, UK.
University of Hull, Hull, UK.
Accid Anal Prev. 2024 Dec;208:107783. doi: 10.1016/j.aap.2024.107783. Epub 2024 Sep 16.
Autonomous vehicles (AVs) offer a range of substantial safety and mobility benefits. However, successful deployment of AVs will involve interacting with other road users, such as pedestrians and other human-driven vehicles. While previous research has focused on factors that influence perceptions of AVs, less work has addressed how specific interactions with other road users influence acceptability from multiple perspectives. Accordingly, we conducted six studies examining how vehicles, either human-driven or autonomous, should behave at a zebra crossing in terms of stopping distance from the crossing, how long a vehicle should wait before setting off, and the influence of traffic context. Using computer-generated videos we found that: the optimal stopping distance was just before the stop line; participants were generally accepting of a vehicle that waited until a pedestrian had fully cleared the crossing before setting off, and sometime earlier; the presence of other vehicles, context and observer viewpoint can affect judgements of vehicle behaviour; autonomous vehicles were judged more harshly than human drivers with learner drivers judged less harshly in some circumstances, and that vehicle size appeared to have little influence over the acceptability of vehicle behaviours. The results are important for informing the design of autonomous vehicle manoeuvres from the viewpoints of vehicle occupants and other road users.
自动驾驶汽车(AV)提供了一系列实质性的安全和机动性优势。然而,要成功部署自动驾驶汽车,就必须与其他道路使用者(如行人和其他人力驾驶车辆)进行交互。虽然之前的研究重点是影响对自动驾驶汽车的看法的因素,但很少有研究涉及与其他道路使用者的具体交互如何从多个角度影响可接受性。因此,我们进行了六项研究,探讨了在斑马线上,无论是人力驾驶还是自动驾驶的车辆,在停车距离、等待时间以及交通环境等方面应该如何表现。我们使用计算机生成的视频发现:最佳停车距离就在停车线前;参与者普遍接受车辆在行人完全穿过路口后再出发的做法,而且有时更早出发也可以接受;其他车辆的存在、环境和观察者视角会影响对车辆行为的判断;自动驾驶汽车比人力驾驶汽车受到更严厉的评判,而在某些情况下,学习驾驶的司机受到的评判则不那么严厉,而且车辆大小似乎对车辆行为的可接受性影响不大。这些结果对于从车辆乘员和其他道路使用者的角度为自动驾驶汽车的操纵设计提供了重要信息。